14 /304
公司避稅與金字塔結構
14
To address the concern, we employ the cash effective tax rate (
CASH_ETR
) as the
second measure of tax avoidance. Following Dyreng et al. (2008, 2010), we define
CASH_ETR
as cash taxes paid divided by pre-tax accounting income in year t. Unlike
GAAP_ETR
,
CASH_ETR
is not biased by changes in tax accounting accruals.
CASH_ETR
reflects avoidance activities that directly affect net income (i.e., permanent book-tax
differences) as well as those that defer income taxes to the later periods (i.e., temporary
book-tax differences). Consistent with prior research, lower values of
CASH_ETR
represent higher levels of tax avoidance (Dyreng et al., 2008, 2010).
One potential issue of the annual
CASH_ETR
can arise from the mismatch between
the timing of cash payment to tax authorities and when the tax is incurred. Cash taxes paid
represent the actual taxes paid by the firm during a given year and could include estimated
tax payments associated with the prior year’s income. In addition, to avoid year-to-year
volatility in annual
CASH_ETR
, our third (fourth) measure of tax avoidance is the long-
term cash effective tax rate following Dyreng et al. (2008). Specifically, we define
LT_
ETR_3Y
(
LT_ETR_5Y
) as the sum of the cash taxes paid over the current year t and the
preceding two (four) years divided by pre-tax accounting income summed over the
corresponding three-year (five-year) period. Prior research suggests that lower values of
LT_ETR_3Y
(
LT_ETR_5Y
) reflect higher levels of tax avoidance (e.g., Gupta and
Newberry, 1997; Rego, 2003; Dyreng et al., 2008).
3.3 Research Design
To examine the association between tax avoidance and the number of layers, the
following empirical model is employed using ordinary least squares (OLS):
AVOID
i,t
=
β
0
+
β
1
LAYER
i,t
+
β
2
SIZE
i,t
+
β
3
MB
i,t
+
β
4
LEV
i,t
+
β
5
ROA
i,t
+
β
6
NOL
i,t
+
β
7
Δ
NOI
i,t
+
β
8
STD_ROA
i,t
+
β
9
Δ
SALES
i,t
+
β
10
CASH
i,t
+
β
11
INTAN
i,t
(1)
+
β
12
RD
i,t
+
β
13
PPE
i,t
+
β
14
ADV
i,t
+
β
15
SGA
i,t
+
β
16
ln
(
NUM_INVESTEE
)
i,t
+
YEARdummy
+
INDUSTRYdummy
+
ε
i,t
Where
AVOID
i,t
is the tax avoidance as measured with
GAAP_ETR
,
CASH_ETR
,
LT_
ETR_3Y
, and
LT_ETR_5Y
.
LAYER
i,t
is the number of layers of the longest investment
chain in firm’s pyramidal structure for firm
i
in period
t
. We take the natural log of
LAYER
i,t
to avoid the non-linearity issue;
SIZE
i,t
the firm size for firm i, year t, measured as
natural log of total assets in year t; market to book ratio
MB
i,t
, measured as the market